LLM Routing on FullStackBench (OOD)
0.8667HVOracle
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Oracle2026.06 | 0.8667 | 0 | 23.45 | |
| MetaRouter2026.06 | 0.6678 | 0.1334 | 22.02 | |
| SW ranking2026.06 | 0.6661 | 0.1434 | 21.66 | |
| Avengers-Pro2026.06 | 0.6548 | 0.1379 | 21.85 | |
| Matrix factorization2026.06 | 0.6385 | 0.1543 | 21.52 | |
| GraphRouter2026.06 | 0.6156 | 0.1656 | 21.8 |